An argument against using "age of first use" as an evaluation measure: A Florida case study

Research has shown that one of the strongest risk factors for substance use problems is early age of first substance use. The longer the initiation of substance use is delayed, the lower the likelihood of problem use in the late teen and early adult years. Consequently the “average age of first use” has become a common measure in evaluating substance use prevention efforts. SAMHSA includes age of first use as one of its national outcome measures (NOMs); and ONDCP requires Drug Free Community (DFC) grantees to report this measure. However, there are several problems with this measure that are discussed in this presentation. While we have been aware of most of these problems for some time, a potentially more serious problem was recently discovered. One county, which has consistently been among the lowest in youth substance use rates, had the youngest age of first use. A detailed analysis revealed that middle school youth had rates and age of first use measures very similar to the state of Florida as a whole, but youth in that county that did not start using before about the 10th grade, were less likely to ever start using – at least before graduation. This resulted in a suppressed average age of first use measure, and led to an erroneous initial assessment that the county was under-performing. A recommendation is made to calculate prevalence of early initiators as an evaluation measure, rather than age of first use. Advantages and operational definitions are discussed.